mutants devoid of N-terminal and HD treated with PNGase F also show difference in their migration pattern in comparison with untreated protein samples indicating that these mutants are also modified with complex glycans.
As described above, zeSho and mSho are attached to the plasma membrane via GPI moiety. So the next question addressed was whether the human homologue is also anchored to the plasma membrane via a GPI moiety. To test this, Sho expressing SH-SY5Y cells were treated with PIPLC. As expected, Sho was found in the cell culture medium after the PIPLC treatment. Sho mutant devoid of the N-terminal domain was also released in to the cell culture medium after the PIPLC treatment, whereas, Sho∆HD was undetectable using either anti-V5 antibody or αSho. But reduced level of Sho∆HD was observed in the cell lysates treated with PIPLC. So the expression of Sho∆HD was further analyzed by non-permeable fluorescence microscopy technique. Indeed, similar to wtSho, expression of Sho∆HD at cell surface was confirmed by fluorescence microscopy.
Collectively, these results confirm that Sho and its deletion mutants are modified with complex glycans, targeted to the plasma membrane through a GPI moiety.
Moreover, deletion of the N-terminal or the HD does not have any detectable impact on Sho biogenesis.
Sho forms homodimers
Many GPI anchored proteins have been shown to form dimers. Dimer formation is not only linked to the physiological function but also to cellular trafficking and targeting to lipid rafts (Mayor & Riezman, 2004; Paladino et al, 2004; Simons & Toomre, 2000). For example, dimerization was demonstrated for CD59, a GPI anchored complement regulatory protein (Hatanaka et al, 1998),
GFRα, urokinase-type plasminogen activator receptor (uPAR/CD87) and CD55 (Airaksinen & Saarma, 2002; Hatanaka et al, 1998). Formation of PrP homodimers have been reported previously and that the internal HD is required for dimerization and is a part of the dimer interface (Meyer et al, 2000a; Priola et al, 1995b;
Rambold et al, 2008b). PrP homodimerization was experimentally shown by introducing a cysteine residue within the HD of PrP (PrPS131C) (Rambold et al, 2008b). A similar approach was previously used to analyze the dimerization of human epidermal growth factor receptor 2 (Erb-2/Her2) and the amyloid precursor protein (APP) (Cao et al, 1992; Munter et al, 2007). Since the HD of Sho is highly homologues to the HD of PrP, it is reasonable to assume that Sho might also be able to form homodimers via its HD.
To address this possibility, a modification was introduced into Sho: the serine residue at 87th position was replaced by a cysteine residue (ShoS87C). If Sho forms a dimer under physiological conditions, the cysteine residues come closer and form a disulfide bond. Under non-reducing conditions the dimers can be detected by a shift in protein migration on SDS/PAGE. Western blot analysis from cultured mammalian cells expressing ShoS87C showed an additional slow migrating band in comparison to wtSho under non-reducing conditions. The slower migrating band disappears in the presence of reducing agents such as β-ME or DDT indicating that Sho forms homodimers and the homodimer is stabilized by an intermolecular disulfide bond via the introduced cysteine residues. PNGase F and PIPLC treatment of ShoS87C revealed that the biogenesis is not altered by the cysteine modification i.e. similar to wtSho, ShoS87C is also complex glycosylated and attached to the plasma membrane through a GPI moiety.
The next question we addressed was whether the N-terminal domain of Sho had an impact on homodimerization. To test this, Sho∆N,S87C (aa 30-56 deleted
in ShoS87C) was expressed in cultured mammalian cells and the protein samples was analyzed by Western blotting under non-reducing conditions. Indeed, similar to ShoS87C, Sho∆N,S87C also forms dimers. These results indicate that the N-terminal domain is not required for Sho homodimerization.
Many membrane proteins form dimers within the cell and are transported to the outer membrane. For example presence of immature GPCR dimers within the ER lumen suggests that dimerization might be an integral part of GPCR maturation and an initial step in the production of functional GPCRs (Milligan, 2004).
Sometimes substrate binding can also induce dimerization of the target proteins at the cell surface. For example brain derived neurotrophic factor (BDNF) binding to TrkB tyrosine kinase receptor induces receptor dimerization (Blum & Konnerth, 2005; Scharfman & McNamara, 2010).
After we obtained experimental evidence for Sho and PrP homodimerization, we further wanted to know if Sho and PrP homodimerization take place within the cell or at the plasma membrane. To test this, Sho and PrP were expressed in mammalian cells in the presence of brefeldin A, a lactone antibiotic that blocks the protein transport from ER to golgi. Western blot analysis revealed that the Sho and PrP homodimers could be detected after the treatment with brefeldin A. This result indicated that Sho and PrP homodimers are formed within the secretory pathway. Further, the effect of glycosylation on homodimerization was investigated. Mammalian cells expressing ShoS87C and PrPS131C were treated with tunicamycin and homodimer formation was analyzed by Western blotting. Indeed, Sho and PrP homodimers can be detected under this condition. Thus, glycosylation is not required for Sho or PrP homodimer formation.
Taken together, Sho and PrP homodimer formation occurs within the secretory pathway and does not require N-linked glycosylation. Since dimerization is a general basic characteristic feature of several membrane receptors, it could be possible that homodimerization of Sho and PrP might be linked with cellular processes such as maturation, cellular trafficking, binding to their receptors, regulation of intracellular signals and endocytosis.
Signal transduction can be initiated between two different cells through protein-protein interaction at the outer surface of the plasma membrane. For example, a homodimer of N-Cadherin, a cell adhesion molecule interacts with another N-Cadherin homodimer of a neighboring cell to form functional cis-trans homotetramer (Kim et al, 2005). Cis homodimers of junctional adhesion molecule 1 (JAM 1) can bind to other cis homodimer of JAM 1 of the adjacent cell and regulate the paracellular permeability and leukocyte transmigration (Kostrewa et al, 2001). Thus, the question arose whether Sho can form trans homodimers at the cell surface. To address this experimentally, two different tags (V5 and HA tags) were introduced into ShoS87C. Cells expressing Sho constructs with two different tags were mixed together and grown until the establishment of cell to cell contacts.
Further, the cells were lysed and the possible formation of a trans homodimer was analyzed by co-immunoprecipitation analysis. Using co-immunoprecipitation analysis, Sho cis homodimer could be readily detected but not trans homodimers or cis-trans homotetramers. However, we cannot rule out the possibility that under certain circumstances Sho trans dimers might be formed in vivo in response to a specific stimulus.
No formation of PrP/Sho mixed dimer
As described earlier, Sho and PrP contain a highly homologues HD. Since, both Sho and PrP can form homodimers via their HD, it is tempting to speculate that Sho and PrP may form heterodimers and HD might be a part of the dimer interface. Using yeast two-hybrid system, it was demonstrated that Sho can interact with PrP, which is mediated by the HD (Jiayu et al, 2009). Further, collision induced dissociation (CID) spectral studies suggested that PrP can be co-purified with Sho or vice versa (Watts et al, 2009). To analyze a possible interaction, PrP along with Sho or PrPS131C with ShoS87C were co-transfected, protein samples were co-immunoprecipitated and analyzed by Western blotting. If Sho and PrP formed a mixed dimer under physiological conditions, introduced cysteine molecules might form a disulfide bridge and different migration patterns for Sho/PrP heterodimer might be seen on Western blot. However, using this approach we were not able to show a PrP/Sho interaction.
The stress-protective activity of Sho
Since Sho and PrP share certain structural features, it seems plausible that the two proteins have also similar physiological functions. Indeed in 2007, Joel Watts and colleagues described a PrP-like neuroprotective activity of Sho. Similar to wtPrP, Sho can protect CGN cells against PrP∆HD-induced toxicity (Watts et al, 2007). In our study we investigated whether this protective activity of Sho is limited to PrP∆HD-induced apoptosis or can also protect cells against other physiological stress agents. To test this, glutamate was employed as a model for excitotoxic stress. Glutamate is considered to be the main mediator of excitotoxicity in the CNS by changing the LTP. Moreover several studies suggested that PrP can ameliorate altered LTP and excitotoxicity induced neuronal
cell death (Collinge et al, 2004; Khosravani et al, 2008; Rambold et al, 2008b;
Whittington et al, 1995).
Mammalian cells expressing PrP or Sho were stressed with acute concentration of glutamate and apoptotic cell death was then measured by staining cells for active caspase-3. Indeed, similar to PrP, Sho can protect SH-SY5Y cells from glutamate stress-induced apoptosis. Further, to map the domains involved in this stress-protective activity of Sho, the deletion mutant clones coding for Sho protein without the N-terminal or HD were prepared. Stress-protective activity of Sho mutants was analyzed by apoptotic cell death as readout. Similar to PrP∆N mutant, N-terminally truncated version of Sho lost its stress-protective activity and lack of HD also makes Sho functionally inefficient. These results indicate that the stress-protective activity of Sho and PrP depend on similar domains i.e, N-terminal domain and internal HD. The impaired stress-protective activity of Sho mutants is not due to improper cellular trafficking, since both the mutants are complex glycosylated and attached to the plasma membrane via a GPI moiety.
With the above results, it is possible to assume that the N-terminal region and the HD are required for the stress-protective function of Sho and PrP. But it is still puzzling as to how the N-terminal and HD are involved in stress-protective signaling? It has been suggested that the intrinsically disordered domains of protein are implicated in protein-protein interaction (Tompa et al, 2009). The N-terminal domain of PrP is intrinsically disordered and by CD spectroscopic analysis of rSho it has been suggested that whole protein might be intrinsically disordered (Watts et al, 2007). Therefore, it could be reasonable to assume that the N-terminal domain of Sho and PrP may interact with their unidentified co-receptors involving intracellular signaling cascade.
Deleting HD in Sho does not lead to neurotoxic species
Experiments in transgenic mice revealed the unexpected finding that by deleting the intrinsic HD, PrP can gain a neurotoxic potential (Baumann et al, 2007; Li et al, 2007a; Shmerling et al, 1998). Interestingly, the neurotoxic potential of PrP∆HD is independent of replication of infectious prion (rev.in (Winklhofer et al, 2008). Co-expression of a single copy of wtPrP can completely abolish the neurotoxic phenotype although the cellular mechanism involved in PrP∆HD mediated toxicity is unknown (Shmerling et al, 1998). Similar to wtPrP, PrP∆HD is also complex glycosylated and targeted to the plasma membrane through a GPI anchor (Winklhofer et al, 2003c). An interesting finding is that similar to PrPC, co-transfection of Sho can counteract PrP∆HD induced toxicity in CGN cells indicating that Sho has PrP-like activity in terms of neutralizing the neurotoxicity induced by PrP∆HD (Watts et al, 2007).
As previously mentioned, sequence similarity between Sho and PrP lies within the HDs. So the next question addressed was, does the removal of the HD from Sho generate neurotoxic species similar to that of PrP∆HD? To test this, Sho∆HD expressing mammalian cells were fixed and stained for active caspase-3 and the apoptosis level was measured. Using a cell culture assay developed in our lab, we could reproduce PrP∆HD-induced toxicity. Mammalian cells expressing PrP∆HD undergo apoptosis and this phenotype is rescued upon co-expression of wtPrP or wtSho. Interestingly, expression of Sho∆HD does not induce apoptosis in cultured mammalian cells indicating that removal of the intrinsic HD of Sho does not result in toxic species at least under the experimental conditions tested. At the same time, mammalian cells undergo apoptosis when Sho∆HD and PrP∆HD are
co-expressed indicating that Sho∆HD does not interfere with PrP∆HD-induced apoptosis. Although Sho and PrP contain identical HD, acquiring a neurotoxic conformer might be exclusive for PrP.
The N-terminal domain of Sho can functionally replace that of PrP
Sho and PrP can protect cells against different toxic insults and their mutants devoid of N-terminal domain (Sho∆N and PrP∆N) are impaired in their stress-protective activity. The next question we addressed was to whether the fusion of the N-terminal domain of Sho with PrP∆N can restore its stress-protective activity. To analyze this, cultured mammalian cells expressing the chimeric protein Sho-PrP (N-terminus (aa 1-63) of Sho fused to PrP∆N (aa 89-251) were stressed with glutamate or co-expressed along with PrP∆HD and apoptotic cell death was measured using activated caspase-3 staining. Indeed, similar to wtPrP, Sho-PrP protects cultured cells against glutamate or PrP∆HD-induced apoptosis.
There is no experimental evidence that either copper or any other metal co-factors bind to the N-terminal domain of Sho. There is no sequence homology between the N-termini of Sho and PrP and the only similarity seems to be that both domains are intrinsically disordered. Based on the hypothesis, that the intrinsically disordered domains can be involved in protein-protein interaction, it is reasonable to assume that N-terminal domain of Sho and PrP bind with same co-receptor to mediate stress-protective signaling.
Sho does not protect cells from PrPSc-induced apoptosis
Formation of PrPSc in the CNS of infected individuals is the crucial event in prion diseases. Apart from PrP, there is no solid evidence of any other protein to play a curtail role in prion disease. After the discovery of Sho, it has been hypothesized that it may have an important role in prion disease. Interestingly, it was reported that Sho is downregulated in prion infected mice. However, Sho mRNA level seems to be unchanged (Lloyd et al, 2009; Watts et al, 2007).
Identifying the SPRN null allele in two patients affected with vCJD also supports the involvement of Sho in prion diseases (Beck et al, 2008).
To address a possible effect of Sho on PrPSc-induced toxicity, a novel method called co-cultivation assay, which is developed in our laboratory, was used. PrP and Sho were expressed in cultured mammalian cells and these cells were then co-cultivated with ScN2a cells, which are constantly secreting PrPSc in the medium. Corroborating previous findings, PrPC expressing cells underwent apoptosis in the presence of PrPSc, whereas in contrast, Sho expressing cells were not affected by the presence of PrPSc. Further-more, we checked whether co-expression of Sho with PrP can block PrPSc-induced apoptosis. To test this, Sho was co-expressed with PrP and these cells were grown in the presence of PrPSc. Co-expression of Sho with PrP failed to protect cells against PrPC dependent PrPSc -induced apoptosis. A recent study in transgenic mice shows that Sho overexpression does not alter scrapie pathogenesis (Wang et al, 2011). Moreover, down regulation of Sho in mice infected with different prion strains has been studied by Kohtaro Miyazawa and Laura Manuelidis and Sho reduction seems to be prion strain specific. For example Sho level seems to be unaffected in mice infected with Asian CJD strain whereas Kuru infected mice show a significant reduction in Sho level (Miyazawa & Manuelidis, 2010).
In summary, our co-cultivation assay results suggest that Sho can neither protect cells from PrPSc-induced apoptosis nor can it mediate PrPSc-induced toxicity. Together with recently published studies, it appears that Sho cannot modulate prion pathogenesis.